No-Multiplier MPPT Monitors Power From Solar Array’s Switch Converter
Fig 1. The multiplierless MPPT system incorporates a buck converter and a control stage. RSHU and the current sense amplifier, U5, sense the system’s output power.
Fig 2. Using a perturb-and-observe algorithm, the system showed a good correlation between the power test program and the current. Total system efficiency was 91%.
Designers working on solar power arrays can simplify maximum power point tracking (MPPT) with a circuit that performs the function without the need for a multiplier. The advantage of this approach is very evident in analog MPPTs.
Everybody would avoid analog multipliers if at all possible. Doing so decreases component count and eliminates the error introduced by the multiplier in the computing of the MPPT. And in a digital circuit you would save one analog-to-digital converter channel and some code.
The principle is simple: The goal is to maximize the power you get from the solar array. For the MPPT to work, it must look at the curve of power versus buck duty cycle. But another curve will do if its maximum coincides with the power curve.
Since the power supplied by the solar array is the same (or at least proportional to) as the power you get from your switching converter (usually a buck converter), you can perform MPPT by looking at the output power. Non-exotic loads (for example, a resistor or a battery to be charged) usually offer voltage and current curves satisfying this requirement. This means the MPPT can work simply by looking at a variable like output voltage or current (see “Achieve MPPT Control Without Power Calculation” at www.electronicdesign.com).
If the application involves charging a battery (a very common task), the best variable to maximize is the current, because the voltage, since it is clamped by the battery, offers a very small range. To transfer the maximum power from the solar panel to the battery, a buck converter is placed between them, controlled by the MPPT. The input of the control stage is only the current flowing to the battery. The output is the buck converter’s pulse-width modulation (PWM) signal.
The MPPT uses a “perturb-and-observe” algorithm. It slightly modifies the PWM value, monitoring the effect on the output current. If it observes a current increase, the MPPT understands it’s doing the right thing and continues modifying the PWM in the same direction. Otherwise, it decides to go in the opposite direction.
Figure 1 shows the MPPT system, comprising the buck converter and the control stage. Connector J1 is not needed for the circuit to work, but makes it easier to program microcontroller U3. (When the microcontroller programmer is not connected, jumpers are needed between pins 1, 3, 5, 7, and 2, 4, 6, 8 of connector J1.)
RSHU and the current sense amplifier, U5, sense the output current. The microcontroller receives this signal, makes the analog-to-digital conversion, and computes the proper PWM signal. The signal is then sent to the MOSFET driver, U1, which drives the buck converter’s pair of power MOSFETs (D8 and D12).
The microcontroller also provides inputs for the temperature and the output voltage measurements, so the circuit can safely charge the battery. U3 also provides UART communications (TX) through pin 2 (TTL_OUT).
The power curve is derived by powering the prototype with a 10-W solar array (with a 21-V unloaded output), loading it with a 6-V lead-acid battery, and running a simple test program (ingvlink) to vary the duty cycle from 0 to 1. The curve has a maximum at 6.2 W, which corresponds to IMPPT = 0.77 A and VMPPT = 8.05 V.
Figure 2 shows the results obtained by feeding the perturb-and-observe algorithm with the power test program and with the current. The two current curves converge at the same value, demonstrating the feasibility of the multiplierless approach. The measured buck efficiency is about 94%, while the entire system efficiency is 91%.
A critical aspect to consider for a solar powered system is that it works under twilight conditions. In fact, microprocessors often suffer from unstable or slow rising edges at power up. For this reason, during testing, we verified that the system worked correctly in the presence of smooth edges on power-up. The critical conditions were simulated at the input by smoothly varying the voltage and the current of a power supply. The system started working properly under these conditions (see “‘Unstable’ Power Supply Simulates Solar Panel Behavior” at www.electronicdesign.com).